LAUSR

S aureus–associated sepsis are urgently required. However, the evidence base to inform such protocols is sorely lacking. Important questions about the epidemiology and clinical treatment of staphylococcal bacteremia and sepsis remain, including optimal choice and duration of antibiotic therapy, benefit from combination therapy or antitoxin therapy, timing of intravenous to oral stepdown, and the role of further investigations such as echocardiography in children. In August 2016, we surveyed Australasian pediatric infectious diseases physicians on the treatment of S aureus infections in children to understand clinical decision making in simple and complicated S aureus infection scenarios. As an example of heterogeneity in treatment approaches, 53% of surveyed clinicians indicated they would add clindamycin empirically in a child with S aureus sepsis. This rose to 66% with confirmed methicillin-resistant S aureus infection. Nonetheless, 80% would feel comfortable randomizing children to a trial of clindamycin vs placebo for toxin-mediated infections, as stated by A. C. Bowen, PhD, via email (November 2016). Schlapbach and MacLaren call for pathogen-specific management bundles for S aureus sepsis. We would argue that empirical clinical evidence should be obtained for the specific aspectsofthesebundlesandsuggest2avenues:first, investigatorinitiated and drug registration trials for S aureus sepsis should include children. There is sufficient overlap in pediatric and adult staphylococcal disease that randomization procedures and often outcome end points can be shared, despite any agedependent differences. Collaborative research is required not merely among pediatricians but between adult and pediatric clinicians. Second, S aureus infection may lend itself to adaptive platform trial approaches.3 Here, patients may be randomized to different interventions in parallel (eg, antitoxin therapy vs placebo or aggressive vs standard source-control). These patients still receive protocolized care, but interventions are randomized where clinical equipoise exists. The ultimate aim is to efficiently determine the best combination of interventions for specific diseasesubgroups.Largesamplesizeswillberequiredforsuchstudies; necessitating multicenter, international collaborative clinical trial networks. Because this is a common and important disease entity, we believe this aim can be achieved.